Image analysis apparatus, image analysis method, and recording medium

Abstract

The present invention reduces the amount of data outputted while maintaining the accuracy of an analysis process with a small delay amount, by an image analysis apparatus provided with: a deduction unit that deduces a second quality concerning an object in a second image, which is different from a first image associated with object data relating to an object to be inputted, on the basis of a first quality concerning the object relating to the object data and on the basis of the state of the object in the second image, the state being obtained by using a state model for deducing the position and the size of the object from the object data, while using a quality model for deducing the second quality concerning the object; and a determination unit that determines whether to use the object data for analysis on the basis of the deduced second quality.

The invention claimed is: 
     
       1. An image analysis apparatus comprising:
 a memory configured to store an instructions; and 
 at least one processor configured to execute the instructions to perform:
 detecting an object and a first state of the object from a first image of a current object data that is input into the image analysis apparatus, 
 deducing, from the first state of the object in the first image, a second state of the object in a second image, input after the first image, by using a state model, 
 predicting, on a basis of the first state and the second state of the object, a first quality regarding the first image and a second quality regarding the second image by using a quality model, and 
 determining not to use the current object data for an analysis when the first quality is lower than the second quality. 
 
 
     
     
       2. The image analysis apparatus according to  claim 1 , wherein
 the state model is a model for deducing a position, a size, and a direction of the object, and 
 the at least one processor is configured to execute the instructions to perform: 
 deducing the second quality, based on at least any of the position, the size, and the direction of the object in the second image. 
 
     
     
       3. The image analysis apparatus according to  claim 1 , wherein
 the at least one processor is configured to execute the instructions to perform: 
 determining to use the current object data for the analysis, when the first quality is higher than the deduced second quality. 
 
     
     
       4. The image analysis apparatus according to  claim 3 , wherein
 the at least one processor is further configured to execute the instructions to perform: 
 determining to use the current object data for the analysis, when the first quality is higher by a predetermined value than a quality related to object data which has been determined to be used for the analysis before determining whether or not to use the current object data being input for the analysis. 
 
     
     
       5. The image analysis apparatus according to  claim 1 , wherein:
 the at least one processor is configured to execute the instructions to perform: 
 deducing the second quality from a state of the object, the state being acquired from the current object data that is the output candidate, and the first quality, by using the quality model, and 
 determining the current object data for use in the analysis, based on the second quality. 
 
     
     
       6. An image analysis method comprising:
 detecting an object and a first state of the object from a first image of current object data input into an image analysis apparatus, 
 deducing, from the first state of the object in the first image, a second state of the object in a second image input after the first image by using a state model, 
 predicting, on a basis of the first state and the second state of the object, a first quality regarding the first image and a second quality regarding the second image by using a quality model, and 
 determining not to use the current object data for an analysis when the first quality is lower than the second quality. 
 
     
     
       7. The image analysis method according to  claim 6 , wherein
 the state model is a model for deducing a position, a size, and a direction of the object, and 
 the second quality is deduced, based on at least any of the position, the size, and the direction of the object in the second image. 
 
     
     
       8. A computer-readable non-transitory recording medium that records a program causing a computer to execute:
 processing of detecting an object and a first state of the object from a first image of current object data that is input into an image analysis apparatus, 
 processing of deducing, from the first state of the object in the first image, a second state of the object in a second image input after the first image by using a state model, 
 processing of predicting, on a basis of the first state and the second state of the object, first quality regarding the first image and a second quality regarding the second image by using a quality model, and 
 processing of determining not to use the current object data for an analysis when the first quality is lower than the second quality. 
 
     
     
       9. The computer-readable non-transitory recording medium according to  claim 8 , wherein
 the state model is a model for deducing a position, a size, and a direction of the object, and 
 the deducing processing deduces the second quality, based on at least any of the position, the size, and the direction of the object in the second image.